Implemental techniques and their effectives for evolutionary algorithms used to solve multilevel lot sizing problems

Multilevel lot sizing (MLLS) problem is a combinational optimization problem which has been proved NP-hard without restrictive assumption on the product structure. Several evolutionary algorithms were developed to solve the MLLS problem in literature, such as genetic algorithm, simulated annealing, swarm particle optimization, soft optimization based on segmentation, ant colony optimization, variable neighbourhood search and so on. In this paper we investigate implemental techniques and their effectives for those evolutionary algorithms used to solve the MLLS problem. Obtained results can be used to specify the characteristics of the solution of the MLLS problems and to help developing more efficient evolutionary algorithms.